Chemoperception of Specific Amino Acids Controls Phytopathogenicity in Pseudomonas syringae pv. tomato

There is substantive evidence that chemotaxis is a key requisite for efficient pathogenesis in plant pathogens. However, information regarding particular bacterial chemoreceptors and the specific plant signal that they sense is scarce. Our work shows that the phytopathogenic bacterium Pseudomonas syringae pv. tomato mediates not only chemotaxis but also the control of pathogenicity through the perception of the plant abundant amino acids Asp and Glu. We describe the specificity of the perception of l- and d-Asp and l-Glu by the PsPto-PscA chemoreceptor and the involvement of this perception in the regulation of pathogenicity-related traits. Moreover, a saturating concentration of d-Asp reduces bacterial virulence, and we therefore propose that ligand-mediated interference of key chemoreceptors may be an alternative strategy to control virulence.

IMPORTANCE There is substantive evidence that chemotaxis is a key requisite for efficient pathogenesis in plant pathogens. However, information regarding particular bacterial chemoreceptors and the specific plant signal that they sense is scarce. Our work shows that the phytopathogenic bacterium Pseudomonas syringae pv. tomato mediates not only chemotaxis but also the control of pathogenicity through the perception of the plant abundant amino acids Asp and Glu. We describe the specificity of the perception of L-and D-Asp and L-Glu by the PsPto-PscA chemoreceptor and the involvement of this perception in the regulation of pathogenicity-related traits. Moreover, a saturating concentration of D-Asp reduces bacterial virulence, and we therefore propose that ligand-mediated interference of key chemoreceptors may be an alternative strategy to control virulence. KEYWORDS chemoreceptors, Pseudomonas syringae, virulence C hemosensory pathways are widely distributed among bacteria and exert a key role in signal transduction processes associated with the response to environmental cues (1)(2)(3). The core of a chemosensory pathway is formed by a complex composed of a chemoreceptor, or methyl-accepting chemotaxis protein (MCP), the CheA histidine kinase, and the CheW adaptor protein. In the canonical pathway, signal binding to the chemoreceptor ligand binding domain (LBD) creates a molecular stimulus that modulates CheA autophosphorylation activity, which in turn alters the transphosphorylation activity of the response regulator CheY. In the case of a chemotaxis pathway, phos-pressure. None of the P. syringae pv. tomato chemoreceptors have been characterized. However, the P. syringae pv. tomato chemoreceptors PSPTO_2480, PSPTO_1061, and PSPTO_2448 ( Fig. 1) are homologous to the amino acid receptors PscA, PscB, and PscC of P. syringae pv. actinidiae (38) and PctA, PctB, and PctC of Pseudomonas aeruginosa (39)(40)(41). dCACHE LBD-containing amino acid chemoreceptors show a wide phylogenetic distribution and were identified, for example, in Halobacterium salinarum (42), Bacillus subtilis (43), Vibrio cholerae (44), Sinorhizobium meliloti (45), and Campylobacter jejuni (46), pointing to an important biological role of this chemoreceptor type (47).
In this work, we have determined the ligand profile of P. syringae pv. tomato PscA (PsPto-PscA) and showed that it exerts a double function, namely, in mediating chemotaxis and in modulating c-di-GMP levels, causing alterations in biofilm development. This receptor was found to play a key role in the infection process, and receptor saturation with its cognate ligands is proposed as an alternative strategy to control virulence.

RESULTS
PsPto-PscA binds L-Asp, D-Asp, and L-Glu. To identify the ligands recognized by PsPto-PscA, the individual LBD was expressed in Escherichia coli and purified from the soluble protein fraction using affinity chromatography. PsPto-PscA-LBD was then submitted to thermal shift assays in which alterations in thermal stability following ligand binding are monitored (48). The method permits the determination of the T m (melting temperature) value, and ligand-induced T m increases of more than 2°C are considered significant. The T m of the ligand-free protein was 37.6°C, and significant T m increases were observed for L-Asp, D-Asp, L-Glu, and N-phthaloyl-L-glutamate, while no T m increase was observed for any of the other proteinogenic amino acids or for D-Ala, D-Asn, D-enantiomers of Asp and L-Glu produced significant exothermic heat changes that diminished as titration proceeded (Fig. 3). Data analysis revealed that L-and D-Asp bound to PsPto-PscA-LBD with the same affinity of 1.2 Ϯ 0.1 M (Table 1), whereas L-Glu bound with a slightly lower affinity (K D [equilibrium dissociation constant] ϭ 3.4 Ϯ 0.1 M). No binding heats were observed for N-phthaloyl-L-glutamate (Fig. 3). Since ITC permits visualization of only high-affinity binding events, it cannot be excluded that this compound binds to the protein with a much lower affinity. We also conducted ITC measurements with several related amino acids that did not cause significant T m shifts (Table 1). In all cases, an absence of binding was noted, indicating that PsPto-PscA-LBD specifically binds L-and D-Asp and L-Glu. These results are in agreement with those observed for the P. syringae pv. actinidiae homologue PscA (38). Finally, microcalorimetric titrations with L-tartrate, a D-Asp homologue abundant in plants, did not reveal binding (see Fig. S1A in the supplemental material).
PsPto-PscA ligands mediate chemotaxis. To investigate the chemotactic response of P. syringae pv. tomato to the three PsPto-PscA ligands, we conducted quantitative capillary chemotaxis assays with the WT (wild-type) strain. Amino acids were used at concentrations ranging from 0.5 to 10 mM. We observed significant responses toward L-and D-Asp and L-Glu, with maxima at 1 mM and 0.5 mM, respectively (Fig. 4). It was reported previously that L-Asp and L-Glu are used by P. syringae pv. tomato as carbon and nitrogen sources (49), a finding that we have confirmed ( Fig. S2A and B). In contrast, D-Asp is not used as a nutrient source (Fig. S2E) by P. syringae pv. tomato and does not alter bacterial growth (Fig. S2F).
To determine whether these chemotactic responses are mediated by PsPto-PscA, we conducted assays with a mutant in which the pscA gene was insertionally inactivated. This mutant did not respond to L-Asp, whereas chemotaxis to D-Asp and L-Glu was  significantly reduced (Fig. 4A), indicating that PsPto-PscA is the sole chemoreceptor for L-Asp, whereas additional receptors are likely to respond to the other two ligands ( Fig. 4B and C). Chemotaxis assays revealed that pscA provided in trans complemented the reduced chemotactic response of the pscA mutant toward L-Asp, D-Asp, and L-Glu (Fig. S3). Taken together, these results indicate that the chemoreceptor PsPto-PscA mediates chemotactic responses to D-Asp, L-Asp, and L-Glu.
Perception of PsPto-PscA ligands controls biofilm formation and swarming motility. The regulatory mechanisms that govern biofilm dynamics are highly complex and remain poorly understood. Several studies highlight the involvement of chemosensory pathways in the biofilm formation process (12,50,51), and a role of specific chemoreceptors in biofilm formation has been reported for Pseudomonas putida  KT2440 and P. aeruginosa (12,52). Furthermore, D-amino acids were found to trigger biofilm disassembly in some bacteria (53,54), while they had no effect on others (55).
To investigate whether PsPto-PscA is involved in biofilm formation, we analyzed WT and mutant strains grown under static conditions during 24 h. A modest but significant increase in biofilm formation was observed for the PsPto-pscA mutant with respect to that of the WT strain (Fig. 5).
To assess the role of PsPto-PscA ligands in biofilm formation, we developed an assay in which saturating ligand concentrations were present, putatively causing complete receptor saturation, preventing a response. The WT strain showed increased biofilm formation in the presence of L-and D-Asp and L-Glu, reaching levels similar to those of the pscA mutant. In contrast, the addition of L-Arg, a compound that does not bind to PsPto-PscA, did not increase biofilm formation in the WT strain (Fig. 5).
Considering the increased biofilm formation of the pscA mutant, and the inverse regulation of biofilm formation and swarming motility in Pseudomonas (35, 56), we assessed bacterial swarming of the WT and mutant strains after 16 h. Inactivation of pscA reduced swarming motility compared to the WT (Fig. 6).
Intracellular levels of c-di-GMP are increased in the PsPto-pscA mutant. The role of c-di-GMP in the transition between motile and sessile lifestyles has been well documented in many bacteria (57)(58)(59) and was also found associated with the inverse regulation of biofilm formation and swarming in species like P. aeruginosa (60,61), Vibrio parahaemolyticus (62), and P. syringae pv. tomato (34). We therefore quantified c-di-GMP levels in the WT and pscA mutant strains by introducing the plasmid pCdrA:: gfp S (Text S1 and Table S1), which harbors a transcriptional fusion of the c-di-GMPresponsive cdrA promoter to a gene encoding green fluorescent protein (63). The  As a positive control, we used the pJBpleD* plasmid, which confers high levels of c-di-GMP (64). The resulting strains were grown on agar plates for 24 h and subsequently analyzed by fluorescence microscopy, showing that the fluorescence intensity was significantly higher in the pscA mutant than in the WT strain ( Fig. 7A to C). The complementation of the mutant with the pscA gene resulted in reduced c-di-GMP levels (Fig. S4). The increased c-di-GMP level in the mutant strain is in accordance with its enhanced biofilm formation (Fig. 5) and the reduced swarming phenotype (Fig. 6). Moreover, the addition of saturating concentrations of PsPto-PscA ligands caused an increase in the c-di-GMP levels in the WT strain on an order similar to that found in the mutant strain ( Fig. 7A to C).
A PsPto-cheA2 mutant is defective in chemotaxis but shows increased c-di-GMP levels. In order to identify the chemotaxis pathway of PsPto-PscA, we constructed a mutant strain of the homologue of P. aeruginosa PAO1 cheA (PA1458), previously described to be involved in the chemotaxis pathway of many MCPs. This gene (PSPTO_1982 or cheA2) is located in P. syringae pv. tomato chemotaxis cluster I (65), and the encoded protein shares 78% sequence identity with P. aeruginosa PAO1 CheA (Fig. S5). In P. aeruginosa, the core proteins of the chemotaxis pathway are encoded by two clusters, namely, clusters I and V. Cluster V contains the CheR and CheV proteins, and this cluster is also found in P. syringae pv. tomato (Fig. S5). The PsPto-cheA2 mutant lost the chemotactic response to the three ligands of PsPto-PscA. This mutant also showed significantly higher c-di-GMP levels than the WT strain (Fig. 8). In order to ascertain whether PsPto-CheA2 is the autokinase involved in chemotaxis, we measured responses to serine, spermidine, and succinic acid. Our results showed that the PsPto-cheA2 mutant lost the ability for chemotaxis to the three compounds tested (Fig. S6).
PsPto-PscA function controls virulence of P. syringae pv. tomato in tomato plants. The decreased swarming motility and increased c-di-GMP levels of the PsPto-pscA mutant may potentially affect virulence. To test this hypothesis, we conducted virulence assays in which leaves of tomato plants were inoculated with the WT and mutant strains. At 6 days postinoculation, bacterial populations were quantified. Data showed significant reductions in both symptom development and bacterial populations for the PsPto-pscA mutant compared to the WT strain. A complemented PsPto-pscA mutant strain restored virulence to WT levels ( Fig. 9A and B).
In subsequent studies, we analyzed bacterial pathogenicity in the presence of the PsPto-PscA ligands. Chemotaxis is based on a ligand gradient covering the chemore-ceptor response range (6), and chemoreceptor saturation with the ligand will prevent taxis. We therefore conducted virulence assays in the presence and absence of saturating concentrations of D-Asp, L-Asp, and L-Glu. To this end, tomato leaves were spray inoculated with a bacterial suspension with or without 1 mM amino acid. As a control, we inoculated plants with bacteria containing 1 mM L-Arg or D-Glu, compounds to which P. syringae pv. tomato does not show chemotaxis (data not shown). After 6 days postinoculation, dramatic reductions in both symptom development and bacterial populations were observed for the plants inoculated with a bacterium-D-Asp mixture  compared to plants inoculated without D-Asp (Fig. 9C and D). In contrast to the drastic decreases in symptom development and leaf colonization that occurred in the presence of D-Asp, only a slight reduction was observed when D-Glu was added to the inoculum. Moreover, the presence of L-Arg, L-Asp, and L-Glu did not alter disease development. As expected, the PsPto-pscA mutant strain was impaired in its virulence despite the addition of L-amino acids to the inoculum (Fig. 9C and D).

DISCUSSION
Knowledge on the function of chemoreceptors in plant-pathogenic bacteria is very scarce. In this work, we have identified the ligands of the PsPto-PscA chemoreceptor and demonstrate that it exerts multiple functions. Apart from mediating chemotaxis, this receptor was shown to be involved in regulating c-di-GMP levels, which was reflected in associated phenotypic manifestations such as changes in biofilm formation or swarming motility. Furthermore, PsPto-PscA was also involved in controlling P. syringae pv. tomato virulence.
There is very substantive evidence that chemotaxis is a key requisite for efficient pathogenesis in plant pathogens (16). It is generally accepted that chemoeffectors released by either plant wounds or stomata induce chemotaxis toward these plant openings, providing access to the apoplast to initiate plant infection. Most of the studies available to date have analyzed mutants in cheA or flagellar genes that resulted in nonchemotactic or nonmotile phenotypes, respectively (16). However, information regarding particular bacterial chemoreceptors and the specific plant signals that they sense is scarce. Here, we show that the PsPto-PscA chemoreceptor binds specifically to D/L-Asp and L-Glu and that signaling through PsPto-PscA affects bacterial virulence in host plants. This ligand specificity is underlined by the observation that L-tartrate, a D-Asp homologue present in plants, failed to bind to PsPto-PscA, and hence, L-tartrate chemotaxis was not altered in the PsPto-pscA mutant (see Fig. S1B in the supplemental material). PsPto-PscA contains a dCACHE LBD; a significant number of dCACHEcontaining chemoreceptors have been reported so far, and many of them are characterized by a broad ligand spectrum since they recognize almost all proteinogenic amino acids. Representative members of this family include the chemoreceptors PctA of P. aeruginosa (39,40) (67). In marked contrast is PsPto-PscA, which has a very narrow ligand range and recognizes only three acidic amino acids. Interestingly, Asp and Glu are the most abundant proteinogenic amino acids in plants (68,69) and specifically in the tomato apoplast (49). Moreover, the presence, although in smaller amounts, of D-amino acids in plants has also been reported (70) and was found to be due to the activity of plant racemases or to the uptake of bacterium-derived amino acids from soil.
It can therefore be hypothesized that the abundance of these compounds in tomato apoplasts has driven the evolution of a chemoreceptor that recognizes these compounds with high specificity and that their detection during plant infection is crucial for optimal infection. Aspartate, next to its abundance in apoplasts, appears to be an amino acid of particular relevance for bacteria since several other chemoreceptors have evolved that recognize this amino acid with high specificity. The Tar receptors of E. coli and Salmonella enterica serovar Typhimurium are the central model chemoreceptors to study chemotactic signaling. Both proteins were found to bind aspartate with high preference, and the obtained dissociation constants are very similar to those obtained in this study for PsPto-PscA (71)(72)(73). Tlp1 (renamed CcaA) of the human pathogen Campylobacter pylori is an aspartate-specific chemoreceptor (46,74) and was found to play an important role in virulence since experimentation with the mutant strain in different hosts resulted in a number of pathological changes in infection experiments (75). Although the corresponding molecular mechanisms remain unclear, these data underline the importance of signaling mediated by aspartate-specific chemoreceptors.
The response to extracytoplasmic signals by bacteria is achieved mainly by two-component signaling systems (TCSs) and chemosensory pathways (76). Many bacteria possess multiple copies of TCSs and chemosensory pathways. Indeed, the genome of P. syringae pv. tomato encodes four chemosignaling pathways (Fig. S7). A central question in signal transduction research resides in establishing whether there is any functional cross talk between these multiple copies of homologous signal transduction systems. Cross talk between signal transduction pathways, while usually considered undesired in the evolution of protein-protein interactions, might be beneficial as a result of an evolutionary response to a given environmental situation. Although for TCSs, significant insight into this issue has been obtained, mainly by the pioneering work of the Laub laboratory (77), much less information is available for chemosensory pathways (12). We show here that mutation of PsPto-pscA increases biofilm formation (Fig. 5) and decreases swarming motility (Fig. 6), which are phenotypes most likely caused by the increase in c-di-GMP levels observed in this mutant (Fig. 7). Furthermore, mutation of PsPto-cheA2 of the chemotaxis pathway generates a dramatic reduction in taxis to several compounds, including the PsPto-PscA ligands, and an increase in c-di-GMP levels. These data suggest that PsPto-CheA2 may interact with other MCPs and that stimulation of PsPto-CheA2 modulates c-di-GMP levels. The results also indicate that the chemotaxis pathway of P. syringae pv. tomato is not an insulated pathway but interacts with different signaling systems, such as the one governing biofilm formation. One of the four chemosensory pathways in P. syringae pv. tomato, encoded by cluster III (Fig. S7), is homologous to the P. aeruginosa Wsp pathway that was shown to modulate c-di-GMP levels in response to as-yet-unidentified environmental cues (10). Data obtained with P. aeruginosa indicate that a given chemoreceptor interacts specifically with only one CheA paralogue (78). In the present case, we hypothesize that PsPto-PscA likely interacts with CheA2, which, however, would signal to its cognate receptor CheY2 as well as to the CheY homologue of the Wsp pathway (WspR_Pspto). CheY2 of P. syringae pv. tomato is a receiver-domain-only response regulator that is likely to interact in its phosphorylated state with the flagellar motor causing taxis. WspR_Pspto is a fusion of a receiver domain with a GGDEF diguanylate cyclase domain. Studies in P. aeruginosa have shown that WspR phosphorylation alters the catalytic activity of its GGDEF domain (79).
Interference of chemotaxis has been proposed to be a strategy to fight bacterial pathogens (80). Chemotaxis is required for localizing to plant openings in order to enter the plant and establish bacterial infection. We show here that saturating PsPto-PscA with D-Asp reduced bacterial infection, which is most likely due to a reduction in chemotaxis toward plant openings. However, the addition of L-Asp and L-Glu did not cause any significant effect on virulence. This may seem to be contradictory at first sight but may be due to the fact that D-Asp cannot be metabolized, whereas both L-enantiomers are efficiently used as nutrients (Fig. S2). Metabolization of these compounds will lead to the formation of a compound gradient, which in turn induces an additional chemotactic response. One may argue that growth promotion by L-Asp and L-Glu may cancel out the negative effect that these ligands had on virulence (as observed with D-Asp). However, L-Arg, a compound that also stimulates P. syringae pv. tomato growth (Fig. S2D) but that is not a PsPto-PscA ligand, did not produce significant effects on virulence. This result indicates that although P. syringae pv. tomato can grow on L-amino acids, the corresponding increase in cell density is rather modest. Moreover, the addition of other D-amino acids like D-Glu, which is not a P. syringae pv. tomato chemoattractant, caused only minor reductions in symptom development and leaf colonization, which is in marked contrast to the severe effects observed in the presence of D-Asp. Taken together, these data show that D-Asp, under saturating conditions, reduces virulence in a specific manner. Therefore, the addition of nonmetabolizable chemoeffectors may be an alternative to inhibit bacterial entry into plants.

MATERIALS AND METHODS
Bacterial strains, culture media, and growth conditions. Bacterial strains and plasmids used in this work are listed in Text S1 and Table S1 in the supplemental material. P. syringae pv. tomato DC3000 and its derivative strains were grown at 28°C in KB (King's B) medium (81). E. coli derivatives were grown at 37°C in LB medium (82). When appropriate, the following antibiotics were added to the medium at the following concentrations: rifampin at 25 g/ml, streptomycin at 50 g/ml, kanamycin at 25 g/ml, ampicillin at 100 g/ml, chloramphenicol at 10 g/ml, nalidixic acid at 10 g/ml, and gentamicin at 5 g/ml.
Identification and classification of MCP signal domains. The search for the MCP signal domains (PF00015) was performed using an ad hoc pipeline as described previously by Río-Álvarez et al. (83). Transmembrane (TM) domains were identified individually using the TMHMM server v. 2.0 (http://www .cbs.dtu.dk/services/TMHMM/) and the DAS transmembrane region prediction algorithm (84). Each MCP was analyzed for Pfam matches (https://pfam.xfam.org/) and categorized according to its LBD. LBDs were considered the domains different from the MCP signal (PF00015) and HAMP (domain present in histidine kinases, adenylyl cyclases, methyl-accepting proteins, and phosphatases) (PF00672).
Construction of the expression plasmid for PsPto-PscA-LBD. A complete list of plasmids used in this study is available in Table S1, and a complete list of primers can be found in Table S2.
The DNA fragment encoding the PSPTO_2480 (amino acids 30 to 278) LBD was amplified by PCR using primers 2480LBDFw and 2480LBDRv (Table S2), and the resulting PCR fragment was finally cloned into the pDEST17 expression vector using Gateway technology (Invitrogen, CA, USA). The verified resulting plasmid, p2480-LBD, was transformed into E. coli BL21(DE3). A detailed description of the construction can be found in Text S1.
Overexpression and purification of PsPto-PscA-LBD. E. coli BL21(DE3) containing p2480-LBD was grown in 2-liter Erlenmeyer flasks containing 400 ml LB medium supplemented with 100 g/ml ampicillin at 30°C. Once the culture reached an optical density at 600 nm (OD 600 ) of 0.5, protein overexpression was induced by the addition of 0.5 mM isopropyl-␤-D-1-thiogalactopyranoside (IPTG). Growth was then continued at 18°C overnight prior to cell harvest by centrifugation at 6,000 ϫ g for 20 min at 4°C. Cell pellets were resuspended in buffer A (30 mM Tris-HCl, 300 mM NaCl, 10 mM imidazole, 10% [vol/vol] glycerol [pH 8.0]) and broken by French press treatment at a gauge pressure of 62.5 lb/in 2 . After centrifugation at 20,000 ϫ g for 1 h, the supernatant was loaded onto a 5-ml HisTrap column (Amersham Biosciences) previously equilibrated with 5 column volumes of buffer A, washed with buffer A containing 40 mM imidazole, and eluted with a linear gradient of 40 to 500 mM imidazole in buffer A. Proteincontaining fractions were pooled and dialyzed into HNG buffer (50 mM HEPES, 300 mM NaCl, 10% [vol/vol] glycerol [pH 8.0]) for immediate analysis.
Thermal shift assay-based high-throughput ligand screening. Thermal shift assays were performed on a Bio-Rad MyIQ2 real-time PCR instrument. Ligands from the compound array (Biolog, Hayward, CA, USA) were dissolved in 50 l of Milli-Q water, which, according to the manufacturer, corresponds to a concentration of 10 to 20 mM. Screening was performed using 96-well plates. Assay mixtures (25 l) contained 15 M protein dialyzed into HNG buffer, SYPRO orange (Life Technologies) at a 5ϫ concentration, and ligands at final concentrations of 1 to 2 mM. In a single well (ligand-free protein), the compound was replaced by water. Samples were heated from 23°C to 85°C at a scan rate of 1°C/min. The protein unfolding curves were obtained by monitoring the changes in SYPRO orange fluorescence. Melting temperatures were determined using the first derivative values from the raw fluorescence data.
Isothermal titration calorimetry binding studies. Experiments were conducted on a VP microcalorimeter (MicroCal, Amherst, MA, USA) at 25°C. PsPto-PscA-LBD was dialyzed overnight against HNG buffer, adjusted to a concentration of 15 M, and placed into the sample cell of the instrument. The protein was titrated by the injection of 8-l aliquots of 0.5 to 1 mM ligand solutions that were prepared in HNG buffer immediately before use. The mean enthalpies measured from the injection of ligands into buffer were subtracted from raw titration data prior to data analysis with the MicroCal version of ORIGIN. Data were fitted with the "one-binding-site" model.
Construction of mutants. The generation of PsPto-pscA and PsPto-cheA2 mutant strains was carried out by single-crossover integration. A detailed description can be found in Text S1.
Quantitative capillary chemotaxis assays. Cultures grown overnight were diluted to an OD 600 of 0.05 in KB buffer and grown at 28°C with orbital shaking. At the early stationary phase of growth, cultures were centrifuged at 1,750 ϫ g for 5 min, and the resulting pellet was washed twice with 10 mM HEPES (pH 7.0). Cells were resuspended in HEPES and adjusted to an OD 600 of 0.25. Next, 230-l samples were placed into each well of a 96-well plate. One-microliter capillaries were filled with the compound to be tested, immersed into the bacterial suspension, and incubated for 30 min. Capillaries were removed from the bacterial suspension and rinsed with sterile water, and the content was expelled into 1 ml of NB medium (1 g yeast extract, 2 g beef extract, 5 g NaCl, and 5 g Bacto peptone [per liter]). Serial dilutions were plated onto NB medium with the appropriate antibiotics, and the number of CFU was determined. In all cases, data were corrected by subtracting the number of cells that swam into buffer-containing capillaries.
Swarming motility assays. P. syringae pv. tomato strains were grown at 28°C for 24 h on KB agar. Cells were resuspended in KB medium to an OD 600 of 1. Five microliters of the bacterial suspension was spotted onto soft KB agar (0.5% [wt/vol] agar). Plates were incubated for 16 h at 28°C at 80% relative humidity (RH) under dark conditions. Swarm colonies were photographed, and the surface area of each colony was quantified using the area selection tool of Adobe Photoshop software with readings in pixels.
Colony-based c-di-GMP reporter assays. Fluorescence intensity analyses using the c-di-GMP biosensor pCdrA::gfp S were carried out according to methods described previously by Corral-Lugo et al. (52), with slight modifications. A detailed description can be found in Text S1.
Tomato virulence assays. P. syringae pv. tomato strains were grown at 28°C for 24 h on KB agar in darkness. Cells were resuspended in 10 mM MgCl 2 and diluted to 10 8 CFU/ml. Three-week-old tomato plants (Solanum lycopersicum cv. Moneymaker) were sprayed with a suspension containing 10 8 CFU/ml. Silwet L-77 was added to the bacterial suspensions at a final concentration of 0.02% (vol/vol). Where indicated, amino acids were added to a final concentration of 1 mM.
Plants were incubated in a growth chamber at 25°C at 60% RH with a daily light period of 12 h. Six days after inoculation, the leaf symptoms were recorded, and bacterial populations from three plants were measured by sampling five 1-cm-diameter leaf disks per plant. The infected leaf disks were washed twice with 10 mM MgCl 2 prior to homogenization to eliminate the bacteria from the leaf surface. Plant material was homogenized in 10 mM MgCl 2 and drop plated onto KB agar supplemented with the appropriate antibiotics. The average number of bacteria per square centimeter isolated from five infected tomato leaves was determined based on log-transformed data.
Statistical analysis. Differences among strains were compared using generalized linear models (GzLMs) when variances were different or one-way analysis of variance (ANOVA) when variances were equal, followed by Fisher's least significant difference (LSD) post hoc test for multiple comparisons, performed using the statistical software package SPSS 22.0 (SPSS Inc., Chicago, IL, USA).